Presence and biodistribution of perfluorooctanoic acid (PFOA) in Paracentrotus lividus highlight its potential application for environmental biomonitoring

The first determination of presence and biodistribution of PFOA in ninety specimens of sea urchin Paracentrotus lividus from two differently contaminated sites along Palermo’s coastline (Sicily) is reported. Analyses were performed on the sea urchins’ coelomic fluids, coelomocytes, gonads or mixed organs, as well as on seawater and Posidonia oceanica leaves samples from the collection sites. PFOA concentration ranged between 1 and 13 ng/L in seawater and between 0 and 794 ng/g in P. oceanica. The analyses carried out on individuals of P. lividus from the least polluted site (A) showed PFOA median values equal to 0 in all the matrices (coelomic fluid, coelomocytes and gonads). Conversely, individuals collected from the most polluted site (B) showed median PFOA concentrations of 21 ng/g in coelomic fluid, 153 ng/g in coelomocytes, and 195 ng/g in gonads. Calculated bioconcentration factors of log10BCF > 3.7 confirmed the very bioaccumulative nature of PFOA. Significant correlations were found between the PFOA concentration of the coelomic fluid versus the total PFOA concentration of the entire sea urchin. PERMANOVA (p = 0.001) end Welch's t-test (p < 0.001) analyses showed a difference between specimens collected from the two sites highlighting the potential application of P. lividus as sentinel species for PFOA biomonitoring.


Matrices collection
Different matrices have been sampled from each one of the 90 specimens of P. lividus (45 per Site) (see Supplementary  P. lividus specimens were brought to the laboratory in a refrigerated ice box immediately after capture and matrix sampling was carried out within the next hour. Isolated matrices were either extracted immediately after their collection or stored at -20 ° C and extracted within 72 hours of their collection.

Materials, equipment and software
LC-MS grade methanol (from Honeywell) was used for extractions (including SPE cartridges preconditioning) and analyses. LC-MS grade water (from PanReac Applichem) was used for HPLC-MS analyses and SPE cartridges preconditioning or washing. Ammonium acetate (from Aldrich) was used as additive for HPLC eluents. Perfluorooctanoic acid analytical standard (>98% from Aldrich) was used for LC-MS calibration curves and to obtain spiking solutions which were freshly prepared and checked for their PFOA content before each batch of analyses. SPE cartridges Strata TM-X-AW (33 µm polymeric weak anion 200 mg/6 mL tubes) were purchased from Phenomenex and used for PFOA extraction from seawater samples. PFOA-free polypropylene micropipette tips were used for quantitative small volume withdrawals. In order to prevent any type of PFOA contamination, the equipment used for sampling and extraction procedures was washed with methanol from the same batch of that used for extraction and analysis. LC-MS analyses were performed using a 6540 UHD Accurate-Mass Q-TOF LC/MS (Agilent Technologies) equipped with a Dual AJS ESI source. Box and jitter graphs have been created using PAST software 3.25, as well as F-test and principal component analysis (PCA) [1]. The Welch's t-test, scatter plots and linear correlations have been realized using Excel 2016 (Microsoft).

PFOA Analyses
Analyses were run in triplicate and were performed on an Infinity 1260 HPLC connected to an ESI-QTOF UHD 6540 MS detector operating in the negative ion-monitoring mode. A volume of 10 μL of sample from extracts or standard solutions was injected in a Poroshell EC-C18 3.0 × 50 mm 2.7 μm column and eluted using a mixture of water, containing 4 mM ammonium acetate, and methanol as eluents with a fixed flow of 0.4 mL/min. The following gradient was used for eluition: from confirming the absence of cross contamination. Quality checks were performed by analyzing a 10 ppb PFOA standard solution every ten samples.

Solid phase extraction of water samples
Briefly, the Strata TM-X-AW SPE cartridges were preconditioned by eluting with 5 mL of methanol followed by 5 mL of water. Then, 500 mL of either seawater or creek's water were passed through the cartridge at a rate of 2 drops per second. The cartridge was then washed with 6 mL of water and allowed to dry. The fraction containing the target analyte was eluted with 10 mL (2 x 5 mL) of methanol and reduced to 1 mL under vacuum.
Spiked samples were prepared by adding 20 µL of an aqueous 1 mg/L stock solution of PFOA to 500 mL of either seawater or creek's water.

P. oceanica analysis
After addition of methanol (5 mL) each P. oceanica sample was sonicated for 20 min at room temperature, centrifuged for 15 min at 3500 rpm, and then the supernatant (1 mL) was taken for HPLC analysis. Spiked samples were prepared by adding 50 µL of an aqueous 1 mg/L stock solution of PFOA to homogenized samples, of 0.5 g each, of P. oceanica.

Coelomocytes and coelomic fluid extraction and analysis
The extraction of PFOA was performed by adding LC-MS grade methanol (5 mL for for samples greater than 0.5 g; 2.5 mL for samples between 0.1 g and 0.5 g) to the sample in a glass vial and sonicating the resulting mixture for 20 min at room temperature. The extraction mixture was then centrifuged for 15 min at 3500 rpm. Then, a portion of the extract supernatant (1 mL) was added to a glass vial to be analyzed by LC-MS. Analyses were run in triplicate and, in case of undetected PFOA, were repeated on a concentrated sample prepared by drying 1 mL of the extract supernatant under a nitrogen current and redissolving the residue in 100 µL of LC-MS grade methanol.
Confirmed undetected PFOA were considered as zero-values in both graphs and statistical analyses.

Sampling campaigns
Three sampling campaigns were carried out at the two sample sites chosen for this study. The main features of the sites and sampling details are summarized in Table S1 below.

Correlation significance
The correlation of log10[PFOA]TOT vs log10[PFOA]CF (Fig. 3 a) had R 2 = 0.75 (thus a correlation coefficient R = 0.86) and was performed on 44 data points (we excluded values where [PFOA] =0 because of its undeterminable log). When the number of observables is equal to 44, the probability to obtain R > 0.7 from two non-correlating variables is p << 0.05% [2]. Therefore, observed correlation is significant. The correlation of log10[PFOA]TOT vs log10[PFOA]CC (Fig. 3 b) had R 2 = 0.46 (thus a correlation coefficient R = 0.68) and was performed on 58 data points (we excluded values where [PFOA] =0 because of its undeterminable log). When the number of observables is equal to 58, the probability to obtain R > 0.45 from two non-correlating variables is p << 0.05% [2]. Therefore, observed correlation is significant.